Publications Board Generative AI Statement

Generative AI in SAA Publications Program 

Statement 5/2024

Introduction:

In January 2024, members of the Publications team received a query about using ChatGPT to generate potential titles for an already completed manuscript draft. This occurred around the same time that American Archivist editors learned that authors had used artificial intelligence (AI) platforms such as ChatGPT in submissions to the Reviews Portal. The Publications Board held a subsequent discussion about the ethical use of AI in Publications Board submissions and/or published content during its February and April 2024 meetings. As ChatGPT and other AI programs are being quickly integrated into creative materials, in addition to the potential of AI to impact archivists’ professional work through tasks such as creating summary documents for finding aids and/or facilitating analysis of large datasets, the Publications Board recognizes the importance of engaging with AI and adjacent topics via scholarship for practitioner training and emerging areas of inquiry or analysis, such as cybersecurity and ethics.¹

Our goal of maintaining the integrity of SAA’s Publications program, which has served archives workers, educators, and students since the 1970s with practical and theoretical literature necessary for managing, preserving, and making accessible collections, is to define standards for both submissions and published work (print and e-publication), and to encourage transparency about any use of generative AI platforms, including ChatGPT, in work submitted.

There is a distinction between using digital support tools for assistance with spelling, transcription, and grammar, and using AI to create original monographs about the archives profession. Multiple disciplines are voicing concerns about generative AI, and we share observations and concerns with peers in myriad roles across the larger publishing industry who include editors, graphic designers, fact checkers, copyeditors, indexers, distribution, marketing, and others—some of whom we contract with for certain aspects of our editorial process.

• First, generative AI may pose substantial issues for copyright and citation integrity. Generative AI is trained on the text of other publications, writings, and images that appear online, a practice that has been challenged multiple times by publishers, authors, and A/V creators.² The corpus of those texts may not be appropriately attributed for copyright or citational purposes, increasing risk of inaccuracies.

• AI-generated content may also reproduce biases or “hallucinate” from existing online content,³ including digital collections,⁴ and contribute to the spread of misinformation.⁵ See, for example, AI legal tools, which cited cases that never existed while creating text for arguments that were intended to be used in a real court of law.⁶

• The speed at which AI-generated content is created does not automatically make that content better in terms of quality. Our peer reviews are concerned with style, voice, and writing form.

To discourage the spread of misinformation and maintain integrity of the publishing program and our responsibility to authors, Society of American Archivists (SAA) members, and readers—some of whom include the general public—the Publications Board is setting forth these Norms and Recommendations for the use of AI in our publications.

Norms and Recommendations

These Norms and Recommendations represent an ethical agreement between authors submitting work to the publishing program, and the Publications Board. They are intended to encourage transparency in the use of AI in the research and publication process. They do not purport to be a legally binding agreement between SAA, the Publications program, or authors submitting manuscripts. However, some points below are relevant to existing requirements in our contracts with authors.

The Board will review these Norms and Recommendations on an annual basis *or following significant changes at the federal or executive level around AI laws or use.

● All submitted manuscript proposals, drafts, and final publications must be conceptualized and written by human beings. The Publications Board’s work, on behalf of SAA, is to reflect in printed and electronic materials theory and technical practice, comprised of observations, lived experiences, analysis, and conclusions made by human beings who have performed or trained others in archives, records management, or cultural heritage-related work. The years of experience it takes to learn that work well enough to summarize it in ways that meet our membership and audience’s needs for book publication are not currently replicable by AI-based programs in ways that reflect the accuracy and standards of our Publications program.

● Authors must disclose the use of AI when a manuscript is submitted. This includes, but is not limited to, AI-generated content in a manuscript, such as large-scale data or metadata analysis. For maximum transparency, this disclosure should appear at the beginning of the manuscript, and at the beginning and throughout the piece where other references or data points may appear.

● All content will go through our standard review channels (board approval and blind peer review, as appropriate).

● Editorial decisions are made based on peer/editorial review, quality of writing, and content.

● Editors reserve the right to request AI transcripts from authors and additional information, including where and how AI was used in the piece.

● Editors reserve the right to provide AI transcripts to peer reviewers as context for the piece.

Adopted 6/5/2024

Next scheduled review 4/2025

NOTES

¹ National Archives. “Inventory of NARA Artificial Intelligence (AI) Use Cases.” September 2023. https://www.archives.gov/data/ai-inventory. 

² United States District Court Southern District of New York, New York Times v. Microsoft Corp, Open AI, Inc., et al. December 27, 2023. https://nytco-assets.nytimes.com/2023/12/NYT_Complaint_Dec2023.pdf. 

³ Noble, Safiya. Algorithms of Oppression. NYU Press, 2018.

⁴ Buolamwini, Joy, Gebru, Timnit. “Gender Shades: Intersectional Accuracy Disparities in Commercial Gender Classification.” Proceedings of Machine Learning Research 81 (2018): 1–15. https://proceedings.mlr.press/v81/buolamwini18a/buolamwini18a.pdf.

⁵ Metz, Cade. “Chatbots May ‘Hallucinate’ More Often Than Many Realize,” New York Times. Nytimes.com. November 6, 2023. https://www.nytimes.com/2023/11/06/technology/chatbots-hallucination-rat....

⁶ Neumeister, Larry. “Lawyers Submit Bogus Case Law Created by ChatGPT. A Judge Fined Them $5000,” Associated Press. apnews.com. June 22, 2023. https://apnews.com/article/artificial-intelligence-chatgpt-fake-case-law....

For further reading

Bowman, Emma. “A College Student Created an App That Can Tell Whether AI Wrote an Essay.” NPR, January 9, 2023. https://www.npr.org/2023/01/09/1147549845/gptzero-ai-chatgpt-edward-tian....

Brainerd, Jeffrey. “As scientists explore AI-written text, journals hammer out policy.” Science magazine. February 20, 2023. https://www.science.org/content/article/scientists-explore-ai-written-te....

Buolamwini, Joy, Gebru, Timnit. “Gender Shades: Intersectional Accuracy Disparities in Commercial Gender Classification.” Proceedings of Machine Learning Research 81 (2018): 1–15. https://proceedings.mlr.press/v81/buolamwini18a/buolamwini18a.pdf.

Chance, Patches. “AI Isn’t Replacing Authors yet, but Exploitation of ChatGPT Is Making Their Jobs Harder.” Paste Magazine, paste.com. March 17, 2023. https://www.pastemagazine.com/tech/ai-art/ai-isnt-replacing-authors-yet-....

Clarke, Neil. “A Concerning Trend – Neil Clarke,” Neil-clarke.com (blog). February 15, 2023. https://neil-clarke.com/a-concerning-trend/.

Harken, Jennifer. “Science Journals Set New Authorship Guidelines for AI-generated Text,” National Institutes of Health. March 2023. https://factor.niehs.nih.gov/2023/3/feature/2-artificial-intelligence-et....

Hillier, Matthew. “Why does Chat GPT generate fake references?” TECHE (blog). Maquarie University. February 2, 2023. https://teche.mq.edu.au/2023/02/why-does-chatgpt-generate-fake-references/.

Leung TI, de Azevedo Cardoso T, Mavragani A, Eysenbach G. “Best Practices for Using AI Tools as an Author, Peer Reviewer, or Editor.” Journal of Medical Internet Research, 2023. Vol. 25:e51584. doi: 10.2196/51584.

Mannheimer, Sara, Doralyn Rossmann, Jason Clark, Yasmeen Shorish, Natalie Bond, Hannah Scates Kettler, Bonnie Sheehey, and Scott W. H. Young. "Introduction to the Responsible AI Special Issue." Journal of eScience Librarianship 13(1): e860.

McElroy, Thad, “How Publishers Can Navigate the AI Revolution,” PublishersWeekly.com. July 15, 2023. https://www.publishersweekly.com/pw/by-topic/digital/content-and-e-books....

Noble, Safiya. Algorithms of Oppression. NYU Press, 2018.

Proceedings of the National Academy of Sciences of the United States. “The PNAS Journals Outlines Their Policies for Chat GPT and Generative AI,” pnas.org. (blog). February 21, 2023. https://www.pnas.org/post/update/pnas-policy-for-chatgpt-generative-ai.

Staiman, Avi. “Guest Post - Academic Publishers Are Missing the Point on ChatGPT - the Scholarly Kitchen.” The Scholarly Kitchen, May 22, 2023. https://scholarlykitchen.sspnet.org/2023/03/31/guest-post-academic-publi....

United States District Court Southern District of New York, New York Times v. Microsoft Corp, Open AI, Inc., et al. December 27, 2023. https://nytco-assets.nytimes.com/2023/12/NYT_Complaint_Dec2023.pdf.

The White House. “Blueprint for an AI Bill of Rights | OSTP | the White House,” November 22, 2023. https://www.whitehouse.gov/ostp/ai-bill-of-rights/.